The Macrotheme Review A multidisciplinary journal of global macro trends

Transcription

1 Hong-Bae Kim*, Taewoo Sohn**, The Macroheme Review 5(4), Winer 016 The Macroheme Review A mulidisciplinary journal of global macro rends Volailiy Transmission beween Convenional Finance and Islamic Finance in Sock Markes 1 Hong-Bae Kim*, Taewoo Sohn** *Division of Business Adminisraion, Dongseo Universiy, , Pusan, Souh Korea ** School of Law, Pusan Naional Universiy. Pusan, Souh Korea, Absrac This paper empirically invesigaed he volailiy spillover effec beween he convenional finance marke and he Islamic finance marke. Our empirical resuls show a unidirecional volailiy spillover from he U.S. convenional sock marke o he Islamic sock indexes of Islamic counries, bu no vice versa. The Islamic sock index (S&P 500 Sharia SHX) of firms lised in he Unied Saes is also influenced by he U.S. convenional sock index (Dow Jones Indusrial Average DJIA). However, our empirical resuls also show here is no volailiy spillover from he U.S. convenional sock index (DJIA) o he Dow Jones Islamic Marke (DJIM) index of globally diversified Islamic socks esablished on Sharia principles. This also holds for he Pakisan sock exchange index. This sudy highlighs he diversificaion benefis of invesing in he various Islamic sock markes in perspecive of porfolio variance. Keywords: Volailiy ransmission; Islamic financial marke; Bi-variae GARCH 1. Inroducion Islamic finance is guided by he principles of Islamic law (Shariah), which prohibis ineres (riba), excessive risk-aking (gharar), and gambling (maysir), and promoes risk-sharing, profisharing, asse-backed financial ransacions, and ehical invesmen. Islamic invesmen has received aenion for wihsanding he recen global financial marke urmoil. If he Islamic finance sysem is decoupled from convenional markes, hen his sysem may provide a cushion agains poenial losses resuling from probable fuure financial crises as well as a srong herapy o global financial crises. The undersanding of how volailiies of and correlaions beween he Islamic and convenional sock markes change over ime including heir direcions (posiive or negaive) and size (sronger or weaker) is of crucial imporance for invesors wih a view o diversifying porfolios o hedge agains unforeseen risks. There is lile lieraure on he volailiy spillover effec beween he convenional finance marke and he Islamic finance marke, and on wheher Islamic finance and convenional finance asse reurns are subsiues or complemens in erms of aking risk. There has been no consensus on he evidence of volailiy spillover beween Islamic sock markes and convenional sock markes 1 This work was suppored by he Naional Research Foundaion of Korea Gran funded by he Korean Governmen(NRF-015S1A5AA ) 1

2 Hong-Bae Kim*, Taewoo Sohn**, The Macroheme Review 5(4), Winer 016 in he prior lieraure. Firs, we refer o he decoupling hypohesis of Islamic equiy finance from convenional equiy finance. An Islamic sock index is argued o be more resilien o financial crisis han is a convenional sock index (Charles, Pop and Darné, 011; Sukmana and Kolid, 01). Second, we se up he rejecion of he decoupling hypohesis of Islamic equiy finance from convenional equiy finance. Ajmi e al. (014) reveal evidence of significan linear and nonlinear causaliy beween he Islamic and convenional sock indexes. They also show poen causaliy beween he Islamic sock marke and financial and risk facors. A sriking resul shows a connecion beween he Islamic sock marke and ineres raes and ineres-bearing securiies, which is inconsisen wih Sharia rules. ÁAlvarez-Díiaz e al. (014) also find ha he Islamic markes are causal and ineracive wih convenional markes and do no perform much beer han he convenional markes during crises. This implies ha he Islamic equiy markes may no be good candidaes for risk diversifiers in asse allocaion or hedgers agains risk exposures. Ajmi e al. (014) show ha he Islamic equiy marke is no isolaed from exernal shocks of differen ypes, regions, and sources. In paricular, he resuls show ha here is a causal link from he Islamic marke o boh he European and Asian convenional sock markes. To explain he impac of convenional marke risk on he Islamic marke, we explore he volailiy spillover effecs beween he Dow Jones Indusrial Average (DJIA) 3 and he Dow Jones Islamic Marke (DJIM) index in he global marke, hose beween he S&P 500 (SPX) and S&P 500 Sharia (SHX) indexes in he U.S. marke, and hose beween he DJIA and several Islamic indexes in Islamic counries, such as he FTSE Bursa Malaysia EMAS Sharia Index (FBMS), he Jakara Sock Exchange Jakara Islamic Index (JAKISL), he Karachi Sock Exchange 100 Index (KSE100) of Pakisan, and he S&P Saudi Arabia Sharia (SPSHSA). Hence, an imporan objecive of his sudy is o invesigae wheher Islamic sock indexes provide more diversificaion benefis han he convenional indexes from he perspecive of volailiy spillover. This requires an empirical invesigaion of he convenional and Islamic sock indexes. We use he bivariae GARCH model o examine volailiy spillovers beween convenional sock markes and Islamic sock markes. This proves ha he selecion of socks based on sric Sharia principles may make some difference in he causal relaionships and links for he DJIM index. Yusof and Majid (007) found he volailiy of ineres rae does no affec ha of he Rashid Hussain Berhad Islamic Index (RHBII), while i does for he Kuala Lumpur Composie Index (KLCI). The KLCI is used o measure he convenional sock marke reurn, while he RHBII is used o measure he Islamic sock marke reurn. 3 The Dow Jones Islamic Marke (DJIM) index was launched in February Since hen, he Dow Jones Islamic indexes have grown o cover 70 counries and 10 regions worldwide. The Dow Jones Islamic indexes exclude companies whose primary business is alcohol, obacco, pork, weapons, enerainmen (e.g., gambling, hoels, pornography) and convenional finance (e.g., banking, insurance).

3 Hong-Bae Kim*, Taewoo Sohn**, The Macroheme Review 5(4), Winer 016. Mehodology In his sudy, we analyze he volailiy and shock ransmission mechanism beween convenional and Islamic sock markes using a bivariae framework of he BEKK parameerizaion (Engle and Kroner, 1995). In his model, he variance-covariance marix of equaions depends on he squares and cross producs of innovaion, which are derived from he following mean equaion:, ~ N 0, H R, (1) 1 where R is he 1 vecor of reurns a ime for each marke, and (ε i, ε ) is he vecor of residuals from he AR(1) sysem (1). Here, we assume ha ε follows a Gaussian disribuion. The n 1 vecor of random errors,, represens he innovaion for each marke a ime wih is corresponding condiional variance-covariance marix H. In order o ensure a posiive semi-definie covariance marix, all elemens mus say posiive during esimaion. The condiional pas marke informaion available a ime 1 is represened by 1. A more pracicable alernaive is he BEKK model given by Engle and Kroner (1995). This model is designed in such a way ha he esimaed covariance marix will be posiive semi-definie, which is a requiremen needed o guaranee non-negaive esimaed variances. There exis numerous parameerizaion mehods for he condiional covariance marix H in Eq. (1). The specificaion we adop is he BEKK model proposed by Engel and Kroner (1995). The BEKK model is sufficienly general and guaranees a posiive definie condiional covariance marix. The sandard BEKK parameerizaion for he bivariae GARCH model is wrien as: H CC A A BH B, () where H = [ h 11, h 1, ], C = [ c 11 c 1 h 1, h, 0 c ], A = [ a 11 a 1 a 1 a ], B = [ b 11 b 1 ], H is a b 1 b marix of condiional variance-covariance a ime, and C is a lower riangular marix wih hree parameers. A is a square marix of coefficiens and measures he exen o which condiional variances are correlaed o pas squared errors. B is a squared marix of coefficiens and shows he exen o which curren levels of condiional variances are relaed o pas condiional variances. h h11, h1, c11 c11 h1, h, c1 c c1 c a a 11 1 a a 1 1, 1, 1 1, 1 1, 1, 1, 1 a a b11 b1 h11, 1 h1, 1 b11 b1, (3) b h 1 b 1, 1 h, 1 b1 b 11, c11 c1 a11 1, 1 a11a1 1, 1, 1 a1, 1 b11h11, 1 b11b 1h1, 1 b1h, 1 (4) 11 1 a a 1 3

4 Hong-Bae Kim*, Taewoo Sohn**, The Macroheme Review 5(4), Winer 016 h, c a1 1, 1 a1a1, 1, 1 a, 1 b1h11, 1 b1bh1, 1 bh, 1 (5) where h 11, denoes he variance of he convenional sock marke reurns, h 1, denoes he covariance of he convenional sock marke reurns and he Islamic sock marke reurns, and h denoes he variance of he Islamic sock marke reurns. of, The significance of diagonal coefficiens 11 h11,, variance 11 a a suggess ha he curren condiional variance h is correlaed wih is own pas squared errors, while he significance of lagged b b indicaes ha he curren condiional variance of 11, h, h is affeced by is own pas condiional variance. In addiion, he significance of he off-diagonal coefficiens a 1 and b 1 indicaes evidence of a volailiy spillover effec from he convenional sock marke o he Islamic sock marke, whereas he significance of off-diagonal coefficiens a 1 and b 1 suggess evidence of a volailiy spillover effec from he Islamic sock marke o he convenional sock marke. We esimae he sysem described by Eq. (1) and () by maximizing he log-likelihood funcion. In he above BEKK model, he off-diagonal parameers in A and B are of paricular imporance in erms of deecing volailiy spillovers across markes. For example, a 1 in A measures he volailiy spillovers in he form of he squared values of he shocks from marke 1 (convenional marke) in he previous period o marke (Islamic marke) in he curren period. Similarly, b 1 in B measures he volailiy spillovers in he form of he condiional volailiy of marke 1 in he previous period o marke in he curren period. We maximize he following likelihood funcion assuming ha errors are normally disribued: L T T (4) 1 T log 0.5 log H 0.5 H 1 1 where θ is he esimaed parameer vecor and T is he number of observaions. Numerical maximizaion echniques were uilized in order o maximize his non-linear log-likelihood funcion. Iniial condiions were obained by performing several iniial ieraions using he simplex algorihm as recommended by Engle and Kroner (1995). The BFGS algorihm was hen used o obain he final esimae of he variance-covariance marix wih corresponding sandard errors. 3. Daa and descripive saisics This sudy analyzes he volailiy spillover effec beween convenional sock markes and Islamic sock markes globally. In his conex, we consider daily closing sock index series for several Islamic sock indexes he DJIM, SHX, FBMS, JAKISL, and SPSHSA as well as for he convenional sock marke indexes he DJIA, SPX, and Korean Composie Sock Price Index (KOSPI00) from January, 00, o November 10, 015. Excluding days when eiher one or boh of he ypes of markes were closed yields a sample size of 3,084 observaions. All sample daa are obained from he Bloomberg daabase. We calculae he coninuously compounded daily reurns by aking he difference in he logarihms of wo consecuive prices, 4

5 Hong-Bae Kim*, Taewoo Sohn**, The Macroheme Review 5(4), Winer 016 ha is, Ri, Pi, Pi, 1 ln 100, where R i, denoes he coninuously compounded percenage reurns for sock indexes i a ime, and P i, denoes he price level of sock indexes i a ime. Tables 1A and 1B presen he descripive saisics and uni roo ess of he daily reurn series for he Islamic and convenional sock markes. In Panel A of Table 1, he JAKISL index presens he highes average reurn, followed by he FBMS index series. Regarding risk, he JAKISL index reurn shows he highes value of sandard deviaion (volailiy). This implies ha he JAKISL index provides higher reurn wih higher risk. Conversely, he FBMS index reurn is found o have he lowes volailiy. Regarding non-normaliy feaures, skewness is negaive for all reurn series. The excess kurosis values for all reurn series are above hree, indicaing he presence of heavy-ailed disribuions and fa ails. Tha is, all reurn series display a lepokuric disribuion wih a higher peak and a heavier ail han he case of a normal disribuion, wih FBMS being he mos skewed and heavy-ailed, implying frequen small gains and exreme large losses. Accordingly, he Jarque-Bera es resuls are consisen wih he aforemenioned deviaions from he Gaussian disribuion, signaling a non-lineariy process. In addiion, he null hypohesis of no serial correlaion is saisically rejeced a he 5% significance level by he Ljung-Box Q es saisic, LB (0), wih a lag of 0 for he squared reurn series. This implies ha he squared reurns exhibi significan signs of serial correlaion. These resuls are in favor of a model ha incorporaes ARCH/GARCH feaures. Table 1A. Descripive saisics DJIA DJIM SPX SHX Panel A: Descripive saisics Mean Max Min Sd. dev Skewness Kurosis *** *** J-B * 106.4*** 701.0* *** LB (0) 389.6*** *** *** *** Panel B: Resuls of uni roo ess ADF *** -5.93*** *** *** PP -60.7*** *** *** *** Noes: J-B and LB (0) refer o he empirical saisics of he Jarque-Bera es for normaliy and he Ljung-Box es for auocorrelaion, respecively. ADF and PP are he empirical saisics of he augmened Dickey-Fuller (1979) and he Phillips-Perron (1988) uni roo saionariy ess, respecively. *** denoes rejecion of he null hypoheses of normaliy, no auocorrelaion, uni roo, non-saionariy, and condiional homoscedasiciy a he 1% significance level. 5

6 Hong-Bae Kim*, Taewoo Sohn**, The Macroheme Review 5(4), Winer 016 Table 1B. Descripive saisics FBMS JAKISL SPSHSA KSE100 KOSPI00 Panel A: Descripive saisics Mean Max Min Sd. dev Skewness Kurosis *** J-B *** Q (0) 09.69*** 436.6*** 49.41*** 1179.*** 064.*** Panel B: Resuls of uni roo ess ADF *** *** *** *** *** PP *** *** *** -5.58*** *** Noes: J-B and LB (0) refer o he empirical saisics of he Jarque-Bera es for normaliy and he Ljung-Box es for auocorrelaion, respecively. ADF and PP are he empirical saisics of he Augmened Dickey-Fuller (1979) and he Phillips-Perron (1988) uni roo saionariy ess, respecively. *** denoes rejecion of he null hypoheses of normaliy, no auocorrelaion, uni roo, non-saionariy, and condiional homoscedasiciy a he 1% significance level. 4. Empirical resuls We proceed wih he esimaion of wo rivariae GARCH models, each conaining he DJIA and DJIM indexes for he global marke, he SPX and SHX in he U.S. marke, and several Islamic indexes (FBMS, JAKISL, KSE100, SPSHSA). The esimaion resuls of he bivariae GARCH model wih BEKK parameerizaion for each variance equaion are repored in Tables and 3 4. In Tables and 3, he symbol h 11, expresses he condiional variance (volailiy) for DJIA a ime, and h 1, shows he condiional covariance beween he DJIA index and he Islamic index in our model. The error erm ε in each model represens he effec of news (i.e., unexpeced shocks) in each model on differen secors. For insance, ε 1, and ε, represen deviaions from he mean due o some unanicipaed even in a paricular marke. The cross values of error erms like ε 1,, and ε,3, represen news in he DJIA index and Islamic indexes in ime period. Tables and 3 show he esimaion resul of he bivariae BEKK (1, 1) model o find he spillover effec beween he convenional sock marke and he Islamic sock marke globally. As menioned earlier, he diagonal elemens in marix A capure he own pas shock effec, while he diagonal elemens in marix B measure he own pas volailiy effec. From Table and 3, he diagonal parameers ( b 11 and b ) in marix B are saisically significan, indicaing he presence of srong GARCH effecs; ha is, own pas volailiy affecs he condiional variance of boh markes. Furhermore, he diagonal parameers ( a 11 and a ) are significan, implying an ARCH effec in boh markes. Firs, all of he diagonal ARCH (a 11 and a ) and GARCH 4 We deeced asymmeric impacs of good news and bad news on volailiy ransmission in he asymmeric rivariae GARCH model, bu here is no significan evidence of asymmeric volailiy ransmission among hese markes. 6

7 Hong-Bae Kim*, Taewoo Sohn**, The Macroheme Review 5(4), Winer 016 (b 11 and b ) parameers are found o be significan a he 1% level a leas, implying a high degree of volailiy clusering in hese series. This indicaes he presence of srong GARCH effecs; ha is, own pas volailiy affecs he condiional variance of boh markes. The offdiagonal elemens of marices A and B capure cross-marke effecs such as shock spillover and volailiy spillover effecs beween he convenional sock marke and Islamic sock markes globally. In order o examine he volailiy spillover effec, we employ he symmeric GARCH (1,1) models based on he BEKK approach. 5 The esimaion resuls of he BEKK model ha includes he convenional indexes DJIA and SPX, and he Islamic indexes relaed o hese indexes, are repored in Table and 3. To check he accuracy of he model specificaions, we employ he Ljung-Box saisic, LB i 0, for squared sandardized residuals. Noe ha he LB i 0 es saisic checks for he serial correlaion of squared sandardized residuals. The insignificance of i LB 0 saisics indicaes he appropriaeness of he GARCH-BEKK model. 5 We deeced asymmeric impac of good news and bad news on volailiy ransmission in he asymmeric bivariae GARCH model, bu here is no significan evidence of asymmeric volailiy ransmission among hose markes. 7

8 Hong-Bae Kim*, Taewoo Sohn**, The Macroheme Review 5(4), Winer 016 Table. Bivariae GARCH model for convenional markes and Islamic markes (I) DJIA DJIM DJIA JAKISL DJIA FBMS DJIA SPSHSA c ***(0.014) 0.148***(0.013) 0.140***(0.013) 0.6***(0.01) c1 0.11***(0.014) (0.039) 0.08*(0.017) 0.018(0.07) c ***(0.006) 0.73***(0.06) 0.101***(0.011) 0.138***(0.033) a11 0.8***(0.047) 0.89***(0.017) 0.79***(0.017) 0.375***(0.03) a (0.048) ***(0.034) 0.053***(0.01) ***(0.048) a1 0.08(0.049) 0.00(0.010) (0.01) 0.059***(0.016) a 0.7***(0.041) 0.7***(0.013) 0.69***(0.016) 0.7***(0.014) b ***(0.013) 0.949***(0.006) 0.951***(0.005) 0.905***(0.01) b1-0.03(0.013) 0.047***(0.013) ***(0.004) 0.059*** (0.017) b (0.013) (0.004) 0.011(0.007) -0.01**(0.006) b 0.976***(0.01) 0.946***(0.005) 0.954***(0.005) 0.949***(0.005) log(l ) Q * * 31.49* Q ** * 93.41**.316 Q * Q * Noes: h 11 denoes he condiional variance for he convenional sock index series, and h is he condiional variance for he Islamic sock index series. The corresponding sandard errors are given in parenheses below each esimaed coefficien. Our esimaed resuls are based on he BEKK parameerizaion. *, **, and *** indicae rejecion a he 10%, 5%, and 1% levels, respecively. We find evidence of a unidirecional shock spillover effec from he DJIA sock marke o he Islamic sock markes excluding he DJIM and KSE100 sock indexes because he coefficien a 1 is significan a he 1% level. 6 Pas shocks in he DJIA marke have a significan effec on he presen volailiy of Eas Asian Islamic markes. Similarly, we idenify a unidirecional volailiy spillover effec from he DJIA marke o he Islamic sock markes excep he DJIM and KSE100 sock indexes owing o he significance of coefficien b 1. Thus, he DJIM global Islamic sock index and he KSE100 of Pakisan sock index are no affeced by convenional sock indexes. 6 I should be noed here ha in assessing he volailiy spillovers, only he significance level of he parameers is imporan, no he sign, because only squared ARCH and GARCH erms ener ino he pahs of volailiy spillovers. 8

9 Hong-Bae Kim*, Taewoo Sohn**, The Macroheme Review 5(4), Winer 016 Table 3. Bivariae GARCH model for convenional markes and Islamic markes (II) DJIA SHX DJIM KOSPI00 DJIA KSE100 SPX SHX c ***(0.01) 0.13***(0.013) 0.133***(0.0130) 0.169***(0.01) c ***(0.015) 0.103***(0.0) (0.050) 0.173***(0.01) c 0.08***(0.004) 0.105***(0.016) 0.381***(0.05) 0.004(0.004) a ***(0.064) 0.97***(0.019) 0.66***(0.016) 0.316*** (0.011) a1 0.4***(0.064) 0.139***(0.030) 0.015(0.014) 0.10 ***(0.008) a *(0.054) **(0.015) -0.00(0.01) *(0.010) a (0.053) 0.19***(0.018) 0.391***(0.00) 0.171***(0.008) b *** (0.0) 0.948***(0.0065) 0.956***(0.005) 0.990***(0.00) b ** (0.01) ***(0.0095) (0.005) 0.010***(0.003) b1 0.01(0.00) 0.005(0.004) 0.006(0.007) ***(0.003) b 0.994***(0.019) 0.974***(0.004) 0.881***(0.01) 0.93***(0.004) log(l ) Q * *** 31.91* 33.75* 1 Q * ** * Q Q Noes: h 11 denoes he condiional variance for he convenional sock index series, and h is he condiional variance for he Islamic sock index series. The corresponding sandard errors are given in parenheses below each esimaed coefficien. Our esimaed resuls are based on he BEKK parameerizaion. *, **, and *** indicae rejecion a he 10%, 5%, and 1% levels, respecively However, here is no significan shock and volailiy spillover effec from he Islamic sock markes o he DJIA marke, excep for SPSHSA. These resuls indicae ha he U.S. convenional sock marke appears o play a more imporan role in influencing he volailiy of he Islamic sock markes. We find evidence of bi-direcional shock and volailiy spillover beween he DJIA marke and he SPSHSA of Saudi Arabia, which ranks firs among oilproducing counries, considering he significance of coefficiens a 1 and a 1, and b 11 and b. In general, our empirical resuls show a unidirecional volailiy spillover from he U.S. convenional sock marke o he sock indexes of Islamic counries esablished on Sharia principles, indicaing ha convenional sock shocks cause he increase of Islamic sock marke volailiy bu no vice versa. The Islamic sock index of firms lised in he Unied Saes (SHX) has also been srongly influenced by he U.S. convenional sock indexes, DJIA and SPX. The 9

10 Hong-Bae Kim*, Taewoo Sohn**, The Macroheme Review 5(4), Winer 016 causal link wih convenional finance markes shows ha here is a sronger Islamic relaionship wih regions in which Islamic finance is more developed. The reverse link is no as srong 7. These findings hus sugges ha he Islamic finance sysem may no provide eiher a good cushion agains financial shocks affecing he convenional markes or large diversificaion benefis for porfolio managers. Paricularly, we find here are unidirecional ransmissions and spillover effecs of shock and volailiy from he convenional equiy sock marke o he Islamic equiy markes of emerging counries such as Indonesia and Malaysia. However, our empirical resuls show here is no volailiy spillover from he U.S. convenional sock index (DJIA) o he globally diversified Islamic sock index esablished on Sharia principle (DJIM). This also holds for Pakisan, which has been isolaed from he global economy. Achsanieal (007) documened ha he inerdependence of he Islamic sock markes ends o be asymmeric in global area. This sudy shows somewha differenly ha here are srong correlaions beween Indonesia and he Unied Saes, Malaysia and he Unied Saes, and Saudi Arabia and he Unied Saes, bu his is no exacly he case across he Islamic sock markes as whole. Majid and Kassim (010) analyzed he marke inegraion among five sock markes, namely, Malaysia, Indonesia, Japan, he Unied Kingdom, and he Unied Saes. They found ha invesors can gain benefis by diversifying in he Islamic sock markes across economic groupings. This sudy conforms o Majid and Kassim s (010) finding ha limied benefis are available if convenional finance invesors only diversify heir invesmens wihin he same Islamic economic groupings. This sudy also highlighs he diversificaion benefis of invesing in he various Islamic sock markes. 5. Conclusion This paper has empirically invesigaed spillover effecs beween he U.S. convenional sock marke and Islamic sock markes esablished on Sharia principles. The esimaion resuls of he bivariae GARCH model are as follows. Firs, we found unidirecional ransmissions and spillover effecs of shock and volailiy from he convenional equiy sock markes o he Islamic equiy markes bu no vice versa. The Islamic sock index of firms lised in he Unied Saes (SHX) has also been srongly influenced by he U.S. convenional sock indexes, ha is, DJIA and SPX. Therefore, we rejec he decoupling hypohesis of he linkage beween Islamic finance markes and he convenional finance marke. Second, our empirical resuls show here is no volailiy spillover from he U.S. convenional sock index (DJIA) o he globally diversified DJIM index esablished on Sharia principles. This resul especially holds for Pakisan, which has been isolaed from he global economy. This sudy invesigaes which Islamic sock index among several Islamic indexes provides more diversificaion benefis compared o he convenional indexes, and finds ha he answer is he globally diversified DJIM index. Overall, he Islamic finance sysem is also exposed o global shocks common o he world financial sysem as well as o conagion risks in he case of economic and financial crises. Therefore, he Islamic sock marke may no be a srong herapy o global financial crises. This evidence leads o he rejecion of he hypohesis of decoupling of he 7 This shows he resricive domain of Islamic invesmens because of he Sharia resricions. Ajmi e al. (014) sae ha convenional markes use several kinds of hedging sraegies agains risks, which may have helped hem somewha o shield hemselves from cross-marke spillovers from he unhedged Islamic marke. Consequenly, he Islamic marke may ouperform is convenional counerpars during bull markes bu underperform in bear markes because of a lack of hedging. 10

11 Hong-Bae Kim*, Taewoo Sohn**, The Macroheme Review 5(4), Winer 016 Islamic markes from heir convenional counerpars, hereby reducing he porfolio benefis from diversificaion wih Sharia-based markes. Hence, furher sudy needs o invesigae he filering crieria for Islamic indexes 8. This requires an empirical invesigaion of he Islamic sock indexes based on he characerisics of he firms and secors based on Sharia principles. References ÁAlvarez-Díiaz, M., Hammoudeh, S., Gupa, R., Deecing predicable non-linear dynamics in Dow Jones Islamic Marke (DJIM) and Dow Jones Indusrial Average (DJIA) indices using nonparameric regressions, Norh American Journal of Economics and Finance 9 (014) 35. Ajmi, A., Hammoudeh, S., Nguyen, D. K., Sarafrazi, S., How srong are he causal relaionships beween Islamic sock markes and convenional financial sysems? Evidence from linear and nonlinear ess, Inernaional Finance Markes, Insiuions and Money 8 (014) Al-Khazali, O., Lean, H., Same, A., Do Islamic sock indexes ouperform convenional sock indexes? A sochasic dominance approach, Pacific-Basin Finance Journal 8 (014) Bollerslev, T., Generalized auoregressive condiional heeroskedasiciy, Journal of Economerics 31(3) (1986) Ciner, C., Gurdgiev, C., Lucey, B. M., Hedges and safe havens: An examinaion of socks, bonds, gold, oil and exchange raes, Inernaional Review of Financial Analysis 9 (013) Dickey, D. A., Fuller, W. A. Disribuion of he esimaors for auoregressive ime series wih a uni roo, Journal of American Saisical Associaion 74(366) (1979) Engle, R. F., Auoregressive condiional heeroscedasiciy wih esimaes of he variance of Unied Kingdom inflaion, Economerica 50(4) (198) Engle, R. F., Dynamic condiional correlaion: A simple class of mulivariae generalized auoregressive condiional heeroskedasiciy models, Journal of Business & Economic Saisics (00). Engle, R. F., Io, T., Lin, W., Meeor showers or hea waves?: Heeroscedasiciy inra-daily volailiy in he foreign exchange markes, Economerica 58 (1990) Engle, R., Kroner, K., Mulivariae simulaneous generalized ARCH, Economeric Reviews, 11 (1995) Engle, R. F., Susmel, R., Common volailiy in inernaional equiy markes, Journal of Business and Economic Saisics 11 (1993) Ho, C. S. F., Rahman, Yusuf, Zamzamin, Performance of global Islamic versus convenional share indices: Inernaional evidence, Pacific-Basin Finance Journal 8 (014) Hood, M., Malik, F., Is gold he bes hedge and a safe haven under changing sock marke volailiy?, Review of Financial Economics () (013) Hosking, J. R. M., The mulivariae pormaneau saisic, Journal of he American Saisical Associaion 75(371) (1980) Among he filering crieria are he concenraion of secors, such as indusrial, echnology, and consumpion services; and he quaniaive crieria of he deb/equiy raio o lower sysemic risks and he limi of ineres-based leverage influenced by ineres rae movemen. 11

12 Hong-Bae Kim*, Taewoo Sohn**, The Macroheme Review 5(4), Winer 016 Inclán, C., Tiao, G. C., Use of cumulaive sums of squares for rerospecive deecion of changes of variance, Journal of he American Saisical Associaion 89(47) (1994) Jawadi F., Jawadi N., Louhichi W., Convenional and Islamic sock price performance: An empirical invesigaion, Inernaional Economics 137 (014) Jihed, M., Mansour, W., Islamic equiy marke inegraion and volailiy spillover beween emerging and US sock markes, Norh American Journal of Economics and Finance (014). Kroner, K. F., Sulan, J., Time-varying disribuions and dynamic hedging wih foreign currency fuures, Journal of Financial and Quaniaive Analysis 8(4) (1993) Kwiakowski, D., Phillips, P. C. B., Schmid, P., Shin, Y., Tesing he null hypohesis of saionariy agains he alernaive of a uni roo: How sure are we ha economic ime series have a uni roo?, Journal of Economerics 54(1 3) (199) McLeod, A. I., Li, W. K., Diagnosic checking of ARMA ime series models using squared residual auocorrelaions, Journal of Time Series Analysis 4(4) (1983) Moghul, U. F., Inroducion o Islamic Finance, Federal Reserve Bank of Boson (006). Narayan, P. K., Narayan S., Modelling he impac of oil prices on Vienam s sock prices, Applied Energy 87(1) (010) Nazrol, K. M., Kamil, Heads we win, ails you lose: Is here equiy in Islamic equiy funds?, Pacific-Basin Finance Journal 8 (014) 7 8. Phillips, P. C. B., Perron, P. Tesing for a uni roo in ime series regression, Biomerica, 75() (1988) Reboredo, J. C., Is gold a safe haven or a hedge for he US dollar? Implicaions for risk managemen, Journal of Banking & Finance 37(8) (013) Rizvi, S. A., Dewandaru, G., Bacha, O. I., Masih, M., An analysis of sock marke efficiency: Developed vs Islamic sock markes using MF-DFA, Physica A 407 (014) Saii, B. The diversificaion benefis from Islamic invesmen during he financial urmoil: The case for he U.S.-based equiy invesors, Borsa Isanbul Review (014). Sansó, A., Aragó, V., Carrion, J. L., Tesing for changes in he uncondiional variance of financial ime series, Revisa de Economía Financiera 4 (004) Shamsuddin, A., Are Dow Jones Islamic equiy indices exposed o ineres rae risk, Economic Modelling 39 (014) Wahyudi, I., Sani, G. A., Inerdependence beween Islamic capial marke and money marke: Evidence from Indonesia, Borsa Isanbul Review 14 (014)